Large Language Models (LLMs) are a novel technology, and their use in education is still developing. Technological challenges, privacy and security concerns, and ongoing issues with the accuracy of their output mean that using them directly to create or evaluate graded assignments is still not advisable.
For ungraded, formative assessments, however, chatbots' ability to provided personalized feedback and react dynamically to student input gives them capabilities no other technology can replicate. What's more, creating customized interfaces for chats with tools like ChatGPT's "GPTs" or Google Gemini's "Gems" can be a surprisingly quick process that doesn't require a ton of development.
Below, we've features a few examples of chatbots built with ChatGPT's features.
This scenario places a student in a situation where they must navigate a conversation. A student journalist is pressing them to reveal confidential information. To successfully complete the scenario, students will need to avoid multiple opportunities in which they can let important information slip.
Try out the scenario and see if you can avoid violating your NDA.
This faculty-facing tool was designed to provide feedback on slide design. Upload or paste a slide or visualization, and it offers feedback based on learning design considerations like accessibility and Richard Mayer's Principles of Multimedia Learning.
Feedback tools like this can offer students guidance on their own assignments, as well.
Jeff Younger, a Clinical Professor of Management Communication, developed this tool in collaboration with the Learning Science Lab to offer students feedback on a writing assignment in one of his courses. It has been customized to offer feedback aligning with the instructor's methods and voice.
In addition, it is not a freestanding resource like the other chatbots on this page. This example was formally integrated into a larger assignment structure where students were asked to incorporate the AI feedback into their revisions and were assessed on multiple drafts throughout their writing process.
This kind of thoughtful deployment of AI tools, where their use is supported and reinforces other parts of the course, can heighten their impact.